The Big Data Paradox

16 Pages·2017·1.69 MB·English

The big data paradox Claver
The Big Data Paradox
Juggling data flows, transparency and secrets
HELVERT
N
A
D/J. V
C
M
O
OT
H
DIT: P
CRE
308
Sprekende kopregel Auteur
The big daTa paradox
On more than one occasion the Militaire Spectator has paid attention to the digital domain,
its characteristics and the role of the military within the digital arena. Recently, articles
related to this subject have been published on military operations, defence policy, cyber
issues, and big data analytics.1 It is abundantly clear that the challenges and opportunities
for the digital era are of great relevance and concern for the military domain. This article
looks into one important aspect of this domain – big data. It will highlight the possibilities
and pitfalls of this phenomenon by putting emphasis on the (neglected) paradox at the
heart of big data developments. It does so from a theoretical, indeed more philosophical,
point of view. Far from lessening its relevancy to military practice, it is argued that such an
approach will help understand the dynamic and complex twenty-first century digital arena
in which military activities play an important and constituent part.2
Dr. A. Claver*
B
ig data is a captivating phenomenon in In other words: big brother will be watching
many respects. Data are to this century you… soon!
what oil was to the previous one: a driver of
growth and change.3 The interconnectedness of The attractiveness of data flows as a powerful
expone ntially growing data flows entails great means of security improvement based upon
promises for personal, commercial, as well as
govern mental use. But is it clear what we mean
* Alexander Claver works at the Dutch Ministry of Defence and is currently following the
by big data? There are many perspectives on big
Executive Master Cyber Security at Leiden University. This article has been written in
data imaginable and at the moment there is no the personal capacity of the author.
single accepted definition. Big data can be 1 See for example Paul A.L. Ducheine, ‘Defensie in het digitale Domein’, in: Militaire
looked upon from a technical, socio­technical as Spectator 186 (4) 2017, p. 152-168. Paul A.L. Ducheine and Kraesten Arnold,
‘Besluitvorming bij cyberoperaties, in: Militaire Spectator 184 (2) 2015, p. 56-70. Erik S.M.
well as governance level, either generic or in
Akerboom, ‘Cyber security. Samenwerken voor een veilige en vitale cybersamenleving’,
detail, and is considered both an opportunity
in: Militaire Spectator 181 (12) 2012, p. 532-536. Allard D. Dijk, Bas Meulendijks and Frans
and a threat. Change holds promises, but is by Absil, ‘Lessons Learned from NATO’s Cyber Defence exercise Locked Shields 2015’, in:
the same token unsettling and intimidating. Militaire Spectator 185 (2) 2016, p. 65-74 and Paul C. van Fenema et al. ‘Big data analytics
From one perspective the advent of big data en Defensie. Visie en aanpak’, in: Militaire Spectator 184 (9) 2015, p. 374-387.
2 In full agreement with Peer H. de Vries (Brig Gen Ret.), who argued in the Militaire
ensures improved transparency from which
Spectator that military practice should be considered from another, philosophical,
society will greatly benefit. From another big
perspective in order to broaden and deepen insights into one’s actions. Peer H. de
data forebodes the horrifying perspective of an Vries, ‘Filosofie voor Militairen’, in: Militaire Spectator 184 (10) 2015, pp. 421-428.
all­knowing and possibly authoritarian, regime. 3 ‘Fuel of the Future. Data is giving rise to a new economy’, in: The Economist, 6 May 2017.
JAARGANG 187 NUMMER 6 – 2018 MILITAIRE SPECTATOR 309
Sprekende kopregel Auteur
CLAVER
(automated) pattern recognition, analysis, and
increased predictive value, is well recognized.
This attractiveness from a security point of view
has aroused suspicion regarding possible
infringements on civil liberties. It is therefore
closely monitored by e.g. human rights activists
and privacy watchdogs. They fear that funda­
mental rights will be jeopardized by the
increased leverage of the state in tracking its
citizens for a range of purposes, e.g. maximizing
tax returns, minimizing social benefit payments,
countering radicalization, or punishing criminal
behaviour. Notwithstanding the validity of these
purposes, these advocates stress that the
position of the individual versus the state has
deteriorated. They plead for more and better
safeguards. This includes transparency and
oversight when it comes to the handling of big
data flows by the state in general, and the
security and intelligence community in
particular.
To counter radicalism, terrorism and other threats in general,
Big data developments: increased transparency Dutch society seems to value a well-functioning intelligence
and secrecy and security apparatus more than before
In 2013 the authors of a short essay cautioned
against what they called the utopian rhetoric of
big data.4 Without denying that big data holds
major potential for the future they claimed that situation or statement that seems impossible, or
the benefits of large dataset analysis were is difficult, to understand because it contains
overstated. To illustrate their point the authors opposite facts or characteristics.6
discussed three, in their opinion understated,
values: i.e. individual privacy, identity, and This article takes a different approach. It does
checks on power. The description of these values not focus on aspects of dichotomy, but
stressed the presence of self­contradictory traits highlights the complementarity and/or
(i.e. paradoxes) in each of the values discussed.5 compatibility of seemingly opposed notions. This
This matches the definition of a paradox as a perspective fits a different definition of a
paradox: ‘A statement that is seemingly
contradictory or opposed to common sense and
yet is perhaps true.’7 This is illustrated by
4 Neil M. Richards and Jonathan H. King, ‘Three Paradoxes of Big data’, in: Stanford Law
Review Online 66, 2013, pp. 41-46. showing that the notions of transparency and
5 Richards and King, ‘Three Paradoxes of Big Data’. 1) the Transparency Paradox, which secrecy do not exclude one another but comprise
concerns the collection of private information by means of big data operations that two sides of the same coin.
are themselves shrouded in secrecy; 2) the Identity Paradox, which emphasizes big
data results operations, but ignores the fact that these techniques seek to identify,
In order to counter radicalism, terrorism and
and therefore work at the expense of individual and collective identity; 3) the Power
Paradox, which deals with the characterization of big data transforming society threats in general, Dutch society at present
without paying attention to the accompanying power effects favouring large seems to value a well­functioning intelligence
government and corporate entities at the expense of ordinary individuals. and security apparatus more than before. This
6 Cambridge Dictionary. See: https://dictionary.cambridge.org/dictionary/english/
apparatus consists of a police force and military,
paradox.
complemented and assisted by the proportionate
7 Merriam Webster Dictionary. See: https://www.merriam-webster.com/dictionary/
paradox. activities (as circumscribed by law) of the two
310 MILITAIRE SPECTATOR JAARGANG 187 NUMMER 6 – 2018
Sprekende kopregel Auteur
The big daTa paradox
The first section of the article briefly points out
the historical roots, definition(s), and main
characteristics of big data, including the
important question of correlation versus
causality surrounding the phenomenon. The
second section deals with the Dutch debate
regarding big data policy and definition. A
conceptual three­layer model of cyberspace is
offered to help structure the intelligence law
discussion by showing that it is predominantly
driven by technical issues (e.g. database design,
intercept possibilities, collection, selection and
search protocols). Attention is put to the fact
that these issues manifest themselves on the
socio­technical level (privacy and security issues).
The third section addresses the paradox of
UYPERS timrapnosrptaarnetn gcoy vaenrdn asneccree ltesv, eliln. Tkhineg f iitn taol stehcet ion
D/V. K offers some concluding remarks.
C
M
O
OT
H
P
big data; some characteristics
Dutch intelligence and security services: AIVD
and MIVD.8 The toolbox of these organizations The first attempts to quantify the growth rate in
naturally includes the full potential of the the volume of data produced have been traced
digital era exemplified in, for example, big data back to the 1940s when the term ‘information
developments. Big data, however, is by definition explosion’ was also introduced.9 Around 1970
connected to a free and transparent flow of computers became inextricably tied to this
information. This does not seem to relate well concept when Gordon E. Moore coined his
with the behaviour of intelligence and security famous, and still valid, rule of thumb that
services. overall processing power for computers will
double every two years (so­called Moore’s Law).10
This article discusses big data developments in The first studies to estimate the amount of new
connection with the simultaneous need for information created annually worldwide
transparency and secrecy. It will zoom in on the appeared in 2000 and 2003. The researchers
concept of big data whose characteristics need to involved (including Hal Varian, now chief
be understood better. Clearer definition, sharper economist at Google) concluded that the amount
demarcation, and the use of conceptual of new information created annually in 1999
modeling will help the current debate wherein amounted to 1.5 billion gigabytes and had
the contributors tend to speak different doubled to 3 billion gigabytes in 2002.11
languages. The article will also show that there
is an apparent, yet not absolute incompatibility
of transparency and secrecy, even though it is
8 AIVD and MIVD are the Dutch acronyms for the civil and military intelligence and
commonly perceived and/or framed as such in
security services. AIVD = Algemene Inlichtingen- en Veiligheidsdienst and MIVD =
the public debate. Recognition and awareness of Militaire Inlichtingen- en Veiligheidsdienst.
this big data paradox will serve current and 9 Gil Press, ‘A Very Short History of Big data’, in: Forbes, 9 May 2013.
future discussions. This is exemplified by the 10 See: http://www.mooreslaw.org/.
11 How Much Information?, School of Information Management and Systems, University
ongoing Dutch debate with regard to a
of California (Berkeley, 2000 and 2003). See: http://groups.ischool.berkeley.edu/
substantial revision of the country’s first
archive/how-much-info/ and http://www2.sims.berkeley.edu/research/projects/
intelligence law of 2002. how-much-info-2003/.
JAARGANG 187 NUMMER 6 – 2018 MILITAIRE SPECTATOR 311
Sprekende kopregel Auteur
CLAVER
‘Big data is high-Volume, high-
Defining big data
There is no definition of big data agreed upon
Velocity and/or high-Variety
yet.12 NASA scientists appear to have coined the
notion first in a paper published in 1997.13 information assets that demand
However, it took the term more than a decade
cost-effective, innovative forms
to become mainstream, and – ultimately – part
of popular culture. The current marketing
of information processing
popularity of big data has little in common with
the original scientific description of the infor­ that enable enhanced insight,
mation revolution, computer accomplishments,
application development (commercial or decision making, and process
otherwise), and the possible implications
automation’ – Doug Laney
connected to this.14 Big data today appeals
above all to the possibility of entering a new
world full of promises, economic opportunities,
and profit.15
A number of current definitions appear to have associated management problems. This led
in common the focus on the magnitude of the industry analyst Doug Laney in 2001 to focus on
amount of data, measured nowadays in Volume, Variety, and Velocity as the key data
thousands of petabytes (1 petabyte = 1,000 management challenges.16 His well­known
terabytes = 1,000,000 gigabytes), and the ‘3Vs’­definition of big data is far from outdated:
‘Big data is high­volume, high­velocity and/or
high­variety information assets that demand
12 Ernst M.H. Hirsch Ballin, et al., ‘Big data in een Vrije en Veilige Samenleving’, cost­effective, innovative forms of information
WRR-rapport, nr. 95 (Amsterdam, Amsterdam University Press, 2016) pp. 33-35. Also: processing that enable enhanced insight,
Dimitri Tokmetzis, ‘Wat is big data?, in: De Correspondent, 11 November 2013. And Gil decision making, and process automation.’17
Press, ‘12 Big data Definitions. What’s Your’s?’, in: Forbes, 3 September 2014.
13 Michael Cox and David Ellsworth, ‘Managing Big data for Scientific Visualization.’ ACM
Other definitions tend to focus less on the
SIGGRAPH, 1 May 1997, 21-38.
14 Though anything but mainstream, the scientific tradition in this respect is not dead. massive amounts of data and more on the
See for an intriguing account of the human and technological limits of computing the opportunities and challenges they offer18 on the
mental exercise by Nick Bostrom, ‘Are You Living in a Computer Simulation’, in: The technical, socio­economic and governance level
Philosophical Quarterly 53 (211) 2003, 243-255.
of cyberspace (see paragraph Modelling Cyberspace
15 ‘Data, data everywhere. Special Report: Managing Information’, in: The Economist,
below.) These definitions point to the impor­
27 February 2010. ‘Fuel of the Future’, in: The Economist. ‘The world’s most valuable
resource is no longer oil, but data’, in: The Economist, 6 May 2017. tance of what can actually be done with the data
16 Doug Laney, ‘3D Data Management: Controlling Data Volume, Velocity, and Variety’, and why its size matters.19 They emphasize the
Application Delivery Strategies 949 (Stamford, META Group, 2001). See: http://blogs. fact that cyberspace data – and the information
gartner.com/doug-laney/files/2012/01/ad949-3D-Data-Management-Controlling-
that can be extracted from it – are giving rise to
Data-Volume-Velocity-and-Variety.pdf. In the next decade Laney continued to work
a new economy.20 This so­called data economy
on his concept and expanded it to ‘12V’s: ‘Deja VVVu: Others Claiming Gartner’s
Construct for Big data’, in: Gartner (January 2012) http://blogs.gartner.com/ derives its strength from self­enforcing network
doug-laney/deja-vvvue-others-claiming-gartners-volume-velocity-variety- effects: ‘using data to attract more users, who
construct-for-big-data/. then generate more data, which help to improve
17 See: http://www.gartner.com/it-glossary/big-data/. services, which attracts more users.’21
18 Hirsch Ballin, et al., ‘Big data in een Vrije en Veilige Samenleving’, p. 33-35.
19 Viktor Mayer-Schönberger and Kenneth Cukier, Big data. A Revolution That Will
Transform How We Live, Work, and Think (London, John Murray Publishers, 2013). Seth Likewise, the Dutch investigative journalist
Stephens-Davidowitz, Everybody Lies: Big data, New Data, and What the Internet Can Dimitri Tokmetzis stresses that his informants
Tell Us About Who We Really Are (New York, Harper Collins Publishers, 2017). are not considering data amounts as such. They
20 Marshall W. van Alstyne, Geoffrey G. Parker and Sangeet Paul Choudary, ‘Pipelines,
refer to big data as connected developments in
Platforms, and the New Rules of Strategy’, in: Harvard Business Review, 94 (4) 2016,
computer technology, consisting of ever more
54-62.
21 ‘Fuel of the Future’, in: The Economist. advanced hardware and software enabling the
312 MILITAIRE SPECTATOR JAARGANG 187 NUMMER 6 – 2018
Sprekende kopregel Auteur
The big daTa paradox
collection of ever more data, and statistics, is not about validating hypotheses, but about
attaching meaning to dispersed data flows by finding interesting links and identifying
relating them to each other.22 patterns that might be relevant. As said, these
analyses might provide unexpected correlations
Correlation versus causality and insights, but run the risk of elevating
Attaching meaning to data by relating, or correlations to causations, even though the
correlating, them to each other touches upon a causality of the linkages found remains uncer­
crucial element of big data and big data usage, tain.27 Ultimately, big data shifts the focus of
one that has not escaped the attention of many inquiry from causation to correlations. Formu­
authors. Distinguishing between correlation and lating a (policy) response will thus depend more
causation is immensely difficult, and often on the knowledge that something is happening
correlation is mistaken for causation. At its core, rather than why it is happening.28
however, a correlation merely quantifies the
statistical relationship between two data points. Some scholars view this positively29 and com­
When one data point changes, the other is likely pare the big data revolution to a classic scientific
to change as well in case of a strong correlation. paradigm shift. According to Rob Kitchin, big
In case of a weak correlation this change is less data analytics enable an entirely new approach
likely to occur. When considering correlations to making sense of the world. Rather than
attention should be paid to the fact that even testing a theory by analysing relevant data, new
strong correlations might occur because of… data analytics seek to gain insights ‘born from
coincidence.23 the data’.30 Jim Gray argues that current data
techniques and technologies are so different that
After all, correlation does not imply causation: it it’s worth distinguishing data­intensive science
only implies probability. Probabilistic outcomes from computational science as a new, fourth
should, therefore, never be taken at face value, paradigm for scientific exploration31 (see
but have to be treated as indications of possible table 1).
outcomes. As a result, any analysis based on
statistical probabilities will, by definition,
produce both false positives (e.g. criminalizing
innocent people) and false negatives (e.g.
allowing security risks to go unnoticed).24 22 Tokmetzis, ‘Wat is big data?’, in: De Correspondent.
23 Viktor Mayer-Schönberger and Kenneth Cukier, Big data. A Revolution That Will Transform
How We Live, Work, and Think (London, John Murray Publishers, 2013) pp. 52-53.
Traditionally, analysis was driven by hypotheses,
24 Hirsch Ballin, et al., ‘Big data in een Vrije en Veilige Samenleving’, p. 38. See also: Dennis
which were validated by collecting and analysing
Broeders, Erik Schrijvers and Ernst Hirsch Ballin, ‘Big data and Security Policies. Serving
data. Insights were extracted from scarce, static, Security, Protecting Freedom’, WRR-Policy Brief no. 6 (The Hague, WRR, 2017) pp. 6-7. See:
and poorly relational data sets with a specific https://english.wrr.nl/topics/big-data-privacy-and-security/documents/policy-
question in mind. Scientific understanding briefs/2017/01/31/big-data-and-security-policies-serving-security-protecting-freedom.
25 Rob Kitchin, ‘Big data, New Epistemologies and Paradigm Shifts’, in: Big data & Society,
today is driven more and more by the (over)
April-June 2014, p. 2. Mayer-Schönberger and Cukier, Big data. A Revolution, p. 70.
abundance of data. When mining these data the
26 See for an interesting view on the lasting importance of small data in the era of big
main challenge will be how to cope with the data developments by Rob Kitchin and Tracey P. Lauriault, ‘Small Data in the era of
variety, messiness, and uncertainty of the big data’, in: GeoJournal 80, 2015, 463-475.
generated data set, bearing in mind that much 27 Hirsch Ballin, et al., ‘Big data in een Vrije en Veilige Samenleving’, p. 38.
28 Kevjn Lim, ‘Big data and Strategic Intelligence’, in: Intelligence and National Security,
of what is collected does not have a specific
31 (4) 2016, p. 622.
question in mind, or is the (unintended) by­
29 Jonathan Shaw, ‘Why Big data is a Big Deal. Information science promises to change
product of another activity.25 the world’, in: Harvard Magazine March-April 2014. See http://harvardmagazine.
com/2014/03/why-big-data-is-a-big-deal.
Here we touch upon an important distinction 30 Rob Kitchin, ‘Big data, new epistemologies and paradigm shifts’, in: Big data & Society,
April-June 2014, p. 2.
between current big data and the infinitely
31 Toney Hey, Stewart Tansley and Kristin Tolle, (2009) ‘Jim Gray on eScience: A
smaller data sets used before.26 Contrary to
transformed scientific method’, The Fourth Paradigm: Data-Intensive Scientific
established scientific practice, big data analysis Discovery (Redmond, Microsoft Research, 2009) p. xviii-xix.
JAARGANG 187 NUMMER 6 – 2018 MILITAIRE SPECTATOR 313
Sprekende kopregel Auteur
CLAVER
Paradigm Form cheaper and more extensive testing of theories,
1. Experimental science Empirical method, describing natural and allowing the continuous assessment of
phenomena theories.33 Nicholas Krohley admits to a wealth
2. Theoretical science Using models, generalizations of data, but speaks of a poverty of insight.
According to him, the ‘fetishization of data’ has
3. Computational science Simulating complex phenomena
led to increasingly complex patterns of
4. Data-intensive science Data-exploration: unifying experiment,
correlation accompanied by increasing failure to
theory, and simulation
contex tualize. He wonders whether an excee­
Table 1 Scientific Paradigm Shifts dingly complex human environment can be
broken down into binary patterns and then
Source: Compiled and adapted from Hey, Tansley and Tolle 2009; Kitchin 2014
reconstructed in a remotely meaningful way?34
Definitions and debates aside, the inevitable
Other scholars are less convinced. Martin Frické conclusion so far must be that the information
argues that so­called data­driven science is a revolution is producing a data­driven society
chimera32. Methodologically speaking, it merely anchored in cyberspace, which will influence
gathers more data, and does not in itself offer people’s lives to a continuously increasing
any explanations or theories, solve scientific extent. For some this is a positive development
problems, or aim to do anything of that nature. heralding great promises.35 Others highlight the
In his eyes, big data encourages passive data negative aspects and warn against harmful
collection, and unsound statistical fiddling. consequences.36
Theory, experimentation, and testing remains
needed as ever. The strength of big data lies,
above all, in supporting this by providing access big data, cyberspace and secrecy:
to (much) larger sample sizes, permitting the dutch case
Digital developments have neither escaped the
32 Wikipedia on Chimera: ‘A monstrous fire-breathing hybrid creature of Lycia (Turkey),
Netherlands nor the attention of the Dutch
composed of the parts of more than one animal.’
government. The economic and societal
33 Martin Frické, ‘Big data and its epistemology’, in: Journal of the Association for
Information Science and Technology, 66 (4) 2015, pp. 651–661. Also Renato Dos Santos, potential of big data (e.g. maximizing tax
‘Big data: Philosophy, Emergence, Crowdledge, and Science Education’, in: Themes in returns, or countering radicalization through
Science & Technology Education, 8 (2) 2015, pp. 115-127. profiling) have been realized as well as the
34 Nicholas Krohley, ‘The Intelligence Cycle is Broken. Here’s How to Fix it’, in: Modern
vulnerabilities with regard to the personal
War Institute at West Point, 24 October 2017. See: https://mwi.usma.edu/intelligence-
sphere (e.g. issues of privacy and equal
cycle-broken-heres-fix/.
35 An outspoken positively inclined author is former Google data analyst Seth treatment). The Dutch government is actively
Stephens-Davidowitz, who published Everybody Lies: Big data, New Data, and What the striving to accomplish a digitalized bureaucracy
Internet Can Tell Us About Who We Really Are. in the foreseeable future. The notion
36 A distinct negatively inclined author is mathematician and former hedge fund data
‘iGovernment’ has become an accepted label in
scientist Cathy O’Neil, who related her experience in Weapons of Math Destruction:
this respect.37 Other clear indications of the
How Big data Increases Inequality and Threatens Democracy (New York, Crown, 2016).
Additional background information can be found in the following interview: Gerard government’s digital awareness are its efforts at
Janssen, ‘Wiskundige Cathy O’Neil en de Weapons of Math Destruction’, in: Vrij formulating big data policy, both in the private
Nederland, 16 November 2016. See: https://www.vn.nl/cathy-oneil-en-weapons- and public sector. The letter to parliament of
math-destruction/.
then Secretary of Economic Affairs Henk Kamp,
37 Corien Prins et al., ‘iGovernment’, in: WRR-Report 86 (Amsterdam, WRR/Amsterdam
published in 2014, has been the point of
University Press, 2011). The WRR-website provides additional information (in Dutch)
on the iGovernment issue including the official government stance. See: http://www. departure with regard to the private sector.38
wrr.nl/publicaties/publicatie/article/ioverheid/. Public sector policy regarding big data has been
38 ‘Kamerbrief over big data en profilering in de private sector. Brief van minister Henk investigated by the Netherlands Scientific
Kamp (EZ) aan de Tweede Kamer over big data en profilering in de private sector, in
Council for Government Policy (Wetenschap­
relatie tot het recht op privacy en het recht op gelijke behandeling’, 19 November
pelijke Raad voor het Regerings beleid or WRR).
2014. See: https://www.rijksoverheid.nl/documenten/kamerstukken/2014/11/19/
kamerbrief-over-big-data-en-profilering-in-de-private-sector. Being an independent advisory body the WRR
314 MILITAIRE SPECTATOR JAARGANG 187 NUMMER 6 – 2018
Sprekende kopregel Auteur
The big daTa paradox
was tasked to advise the government on this
matter, which resulted in the publication of
several reports in 2016.39
The WRR also looked into related cyber matters
within the project Freedom and security in the cyber
domain.40 This resulted in a number of publica­
tions advocating the state’s responsibility for the
‘public core of the internet’. States need to
involve themselves by making sure that the
internet core – i.e. the central protocols and
infrastructure considered to be public good – are
safeguarded from state interference. The project
emphasized the interconnectedness of technical,
socio­technical and governance elements in the
cyber domain and stressed that cyber policy
issues are, by necessity, played out intern ationally
and cannot be confined to the national level (for
national security considerations) or left to market
forces alone. The council, therefore, aimed ‘to D
pcinor ohwvehirdeicenh tk ntfhoorewe iilngetnde grpeeo stltoisc ayos ffs oeisrct ot ihnneo dmceyivbce,e lrpo hpdiyonsmgic aaai ln a, nodn e O RIJKSOVERHEI
national security, on the one hand, and political HOT
P
and economic freedom, on the other, are weighed People are sharing more and more data in the digital domain by social media
up against one another.’41
A related topic within the current Dutch public data is described as follows: ‘…the phenomenon
debate is the new intelligence law.42 Within this that manifests itself among others in the fact
debate the earlier mentioned concepts of big data, that the amount of data is growing expo­
cyberspace, transparency and secrecy – and by nentially, data collections are becoming bigger
proxy, freedom and security – are linked and and more complex as a result of which relevant
hotly contested. The inability so far to find data can no longer be stored physically or
common ground owes much to the failure of logically in a location or in a system....’ 43
clearly demarcating and/or defining the issue(s) at
stake. Two examples will suffice to illustrate this.
39 Hirsch Ballin, et al., ‘Big data in een Vrije en Veilige Samenleving’. An English
translation of the aforementioned report is: Broeders, et al., ‘Big data and Security
Defining big data in the Netherlands
Policies’.
The previous section on big data has clearly
40 Dennis Broeders et al. ‘De Publieke Kern van het Internet. Naar Buitenlands
shown the elusiveness of the notion. Notwith­ Internetbeleid’, in: WRR-rapport nr. 94 (Amsterdam, Amsterdam University Press,
standing Doug Laney’s clear and concise 2015). English translation: Dennis Broeders, ‘The Public Core of the Internet. An
‘3Vs’­definition of big data, no communis opinio International Agenda for Internet Governance’, WRR-Policy Brief no. 2 (The Hague,
WRR, 2015). See: https://english.wrr.nl/publications/reports/2015/10/01/the-public-
on the subject exists to date. The arduous
core-of-the-internet.
attempt of the Dutch government to clarify the
41 Broeders et al. ‘De Publieke Kern van het Internet.’
issue in relation to the revision of the intelli­ 42 See https://zoek.officielebekendmakingen.nl/dossier/34588. See also: https://www.
gence law merely confirms the fuzziness of the internetconsultatie.nl/wiv/details.
concept and the difficulty of demarcating it. 43 Tweede Kamer der Staten-Generaal 2016-2017, ‘Regels met betrekking tot de
inlichtingen- en veiligheidsdiensten alsmede wijziging van enkele wetten (Wet op de
inlichtingen- en veiligheidsdiensten 20..)’, in: Memorie van Toelichting, Kamerstuk
Within the Memorie van Toelichting (Explanatory
34588-3, p. 130. See: https://zoek.officielebekendmakingen.nl/kst-34588-17.html
Notes) concerning the new intelligence law big (author’s translation).
JAARGANG 187 NUMMER 6 – 2018 MILITAIRE SPECTATOR 315
Sprekende kopregel Auteur
CLAVER
large structured and unstructured data from different Modelling cyberspace
Data
sources Big data developments are inextricably connec­
data-driven, automated searches for correlations, in ted to cyberspace. But, most people are unable to
particular with the potential for analysis of the present answer basic questions, such as: What is
Analysis
(real-time analysis) and the future (predictive analysis) cyberspace? How is cyberspace being governed?
Who are its attackers and what are their
analysis should result in actionable knowledge, to motives? How does the (underlying) technology
Actionable
be made applicable for decision-making at group or work?, etc.47 It stands to reason that without the
Knowledge
individual level. existence of generally accepted answers,
Table 2 Big Data Characteristics (WRR) clear­cut definitions, and suitable demarcations,
it becomes difficult to see eye to eye with each
Source: Hirsch Ballin et al. 2016
other when perceptions and/or interpretations
differ.
This wording – ‘hidden’ as a subordinate clause
on page 130 – is anything but exact. It does A conceptualization of cyberspace is, therefore,
not constitute a clear­cut definition and is urgently needed as will become clear from the
contrasted on the very same page by referring debate in the Netherlands regarding the new
to a characterization of big data in a WRR­ (revised) intelligence law (see paragraph Debating
report. The authors of this report hold that the Secrecy below). A promising start in this respect
concept of big data is ambiguous. Instead of has been the approach of Van den Berg et al. In
providing a definition, they therefore chose to an award­winning paper, published in 2014, the
focus on what they consider the three main authors suggest a conceptual model dividing
characteristics of big data: data, analysis, and cyberspace into twelve cyber subdomains,
actionable knowledge44 (see table 2). arguing that these domains need to be analyzed
on three separate, but interconnected layers: a
Big data here is not seen as a well­defined – or technical, socio­technical, and governance
even a definable – concept, but as the dynamic layer48 (see figure 1).
interplay between the three displayed charac­
teristics. According to the authors, this leaves From the model follows that the traditional
room to discuss the use of data analysis in public inclination to concentrate on and investigate the
policy making.45 This characterization is technical aspects of cyberspace does not suffice.
subsequently accepted in the Memorie van It is imperative that socio­technical and
Toelichting with the concluding remark that an governance aspects are considered as well.
interpretation of big data as provided by the Historically, the technical layer focusing on
WRR is in line with the assumptions of the robust communication services and information
proposed law revision.46 security has received the most attention.
However, global interconnectivity and huge
numbers of applications with an easy to use
human interface have given rise to a socio­
technical layer. Here, people perform a vast
44 Ernst Hirsch Ballin, et al., ‘Big data in een Vrije en Veilige Samenleving’, pp. 33-35. range of cyber activities, which translates into
[Author’s translation]. the complex interaction of billions of people
45 Broeders, Schrijvers and Hirsch Ballin, ‘Big data and Security Policies’, p. 6.
active in cyberspace with the available IT­
46 Tweede Kamer der Staten-Generaal 2016-2017, ‘Regels met betrekking tot de
systems – i.e. data storing and data processing
inlichtingen- en veiligheidsdiensten alsmede wijziging van enkele wetten’, in:
Memorie van Toelichting Kamerstuk 34588-3, p. 130. systems, including to an increasing extent
47 Jan van den Berg, et al., ‘On ( the Emergence of ) Cyber Security Science and its intelligent and autonomous decision­making
Challenges for Cyber Security Education’, in: NATO STO/IST-122 and Cyber Security systems. The governance layer consists of the
Academy (Den Haag, 2014) See: https://www.csacademy.nl/images/MP-IST-122-12-
large and complex number of human actors and
paper-published.pdf, p. 1.
organizations that govern both the technical and
48 Van den Berg, ‘On ( the Emergence of ) Cyber Security Science and its Challenges, p.2.
49 Ibidem. socio­technical layers.49
316 MILITAIRE SPECTATOR JAARGANG 187 NUMMER 6 – 2018
Sprekende kopregel Auteur
The big daTa paradox
Debating secrecy
Tdtdeeaocbt haaa ntsceiuoc bmianslm tptahonuestn siNiaicblea itdlthiietoeginerrsl eai enon fttd ehlsrae crc egienpentt­tseieoclrlanislg eaae nrnaoducu ettno hlmdae w tarh tiseekd s nsportation F ood (chaiGno) v e rCnhaenm cineidcauls&trNyu c l e a r D rsienckting w
ipAnrdvihvoaelvcreiynd vg ei ntros a uiptsps s r‘eiocGvuoarvli etoyrfn asmurgceuhnm tm’e penrtthi nioscd iapst.l epTslh ateyh eh ere. m Tra S ocio-technical or ater Ene
Dpruotpcohs galo vuepr nfomr eonntl ihnaes c pounts uitlst aintitoenll.i5g0e Anccceo lradwin g eleco Technical rgy se
T c
tino t1h,1e1 M4 irneissptoryn soefs I, netveernnlayl dAifvfiadierds tbheitsw reeesnu lted tor F
confidential responses and responses open to re ani
d n
public scrutiny.51 ro lag es laic
Twpaoonvfohi ettdiehser ssscn ufuiotgoeisbhratsje lt ei (i cncig(nbtocne onur dtssarh eerntr oqedevd uxoci ec­rbmaeniytns ci o,tcce ertiossed mamwcenhrpe,d)n ar n reineeec xiexcae­aeltps gi,itov seocsesro tudml)oy. e5pp as2dae t aetArrnebta irandtcsit uuo etomhlndafe rbi re r eL redroy tceilfbaSu P& n o i t acriltbsiunPim d a t n e m reegtaaWn a m rotces htl aeH rotc
attention:
1. Large­scale interception of cable Figure 1 A Conceptualization of cyberspace in layers and (cyber) subdomains
communication;
Source: Van den Berg et al. 2014
2. Search through large amounts of data;
3. Automated access and analysis of databases; lesser extent on socio­technical layers, as can be
4. Obligating companies and organizations to seen from other contributions. For under­
decrypt communication. standable reasons the technically possible
interception, collection and storage of huge
Criticism on these issues boiled down to: amounts of data tickles people’s imagination.
Catchphrases such as ‘select before you collect’,
1. Matters of necessity, proportionality, and ‘collect before you select’ and even ‘select while
subsidiarity; you collect’ exemplify the main road taken by
2. Question marks concerning privacy goals; most researchers.53 A look at the reports of the
3. The technical impossibility of compliance.
This emphasizes the importance of a well­
50 See: https://zoek.officielebekendmakingen.nl/dossier/34588 and https://www.
functioning oversight mechanism, given the fact
internetconsultatie.nl/wiv/details.
that intelligence and security services already 51 Maurits Martijn, ‘Wat zijn de wensen van dit kabinet voor de geheime diensten?’, in:
possess far­ranging powers by law regardless of De Correspondent, 11 January 2016 https://decorrespondent.nl/1632/wat-zijn-de-
the actual outcome of the intelligence law wensen-van-dit-kabinet-voor-de-geheime-diensten/50193792-ce33fa45.
52 https://www.internetconsultatie.nl/wiv/details and Maurits Martijn, ‘Vier redenen
revision. Though governance encompasses more
waarom de nieuwe aftapwet een slecht idee is’, in: De Correspondent, 12 July 2017
than oversight, the third layer within the cyber
https://decorrespondent.nl/7054/vier-redenen-waarom-de-nieuwe-aftapwet-een-
domain (see figure 1) has at long last appeared slecht-idee-is/713051614880-1a2bce5c. NB: This article is a later version of the article
on the horizon as an integral part of an mentioned in note 44 by the same author. The two interactive articles link to many
indispensable system of checks and balances. important contributions concerning the intelligence law debate.
53 Bart Jacobs, ‘Select before you collect’, in: Ars Aequi, Vol. 54 (No. 12) pp. 1006-1009,
2005 and Bart Jacobs, ‘Select while you collect. Over de voorgestelde
The intelligence law debate, however, has
interceptiebevoegdheden voor inlichtingen- en veiligheidsdiensten’, in: Nederlands
remained focused on the technical and to a Juristenblad Vol. 91 (Den Haag, 29 January 2016) p. 256-261.
JAARGANG 187 NUMMER 6 – 2018 MILITAIRE SPECTATOR 317